Power amp chip??

[Jaicen_solo]

Slightly O/T I know, but does anybody have any recommendations for a h
Medium power amplifier chip for a solid state amp??
I'd like about 20w, to make a twin stereo amp with a stomp box front end.
Anybody have any recommendations??

[Johan]

Lm1875

[mth5044]

TDA2005R can do stereo, I believe. I'm working up an amp design with it, but not in its stereo application. Check out the datasheet for circuit examples.

[Jaicen_solo]

Thanks for the recommendations gents. The TDA package looks great, but it's 2x10w. I need at least 2x25w to keep up with the volume levels required. It needs to be as loud as a cleanish fender deluxe reverb, so 2x25w should work shouldn't it?? I'm pretty sure that 2x50w does not equal 50w. Anybody that had built one, can you give me some idea if this would work driving 2x12" celestions?

Any recommendations for a front end? I'm thinking a fet based ROG type thing, perhaps using a transformer phase splitter before the power amp???

[R.G.]

If I may make my standard suggestion - 

1. Work out the necessary power supply, heat sink, and enclosure designs FIRST before you go haring off looking for power amp chips or circxuits. To a first approximation, any power amp chip or circuit that's capable of the power will be fine if it's implemented well. The power supply design and parts will limit you much more than what circuit you use. Figure out the power supply, heatsinks, and enclosure FIRST, or you will get to build it two times, not once.

2. It is really very difficult to beat the LM3778 or some of its later relatives for a power amp. The LM3778 costs about $5.00 each for a reliable (if properly powered and cooled) circuit that is well tested in the market. The difference in price between a couple of LM3778s and smaller dual-ampliifer solutions or other power amp chips vanishes into the rounding error of the power supply, heat sink, and enclosure costs

3. Wiring up the AC power side of things for an amplifier is *literally* betting your skills against your life and the life of everyone in the building where it's used. Be SURE you know what you're doing.

There are several variations on this advice that I've written up in this forum before. You might want to search for and read some of those posts for other perspectives on the same advice.

[R.G.]

It strikes me that I should write up a coherent set of this advice with some tables of alternatives and some example calculations and put them up at GEOFEX so I could just point to them. I'll go do that.

[mth5044]

Oops, I had forgot that I was running it differently for different watts! I appologize. There are many variations of the TDA chips out there, I'm sure there is one that suits what you need.

As far as getting a solid state amp to be as loud as a tube amp, I think there is more to it than matching watts. Someone smarter than I will need to explain it, but based of the amps I've played through, it really depends a lot on the speakers, efficiency, etc more than the stated watts. I've also played tube amps that were lower wattage than SS amps, but were 'perceived' as being louder.

[FiveseveN]

As far as getting a solid state amp to be as loud as a tube amp, I think there is more to it than matching watts.

Yeah, make them distort. But not in the abrupt fashion that abundant NFB conduces.

[Jaicen_solo]

R.G, as always thanks for the advice.
The power supply etc will be as big as the circuit demands, hence asking for opinions before buying any parts.
Let me also make it clear that this is going to be designed and built to be as close to bulletproof as is humanly possible. Hence,
Heat sinking overkill, and PT's that are double the required current etc.

The wiring in something like this is nothing to frighten me, I've built tube stuff and wired household stuff so it should be fine.

[R.G.]

Good. You're past the first hurdles.

My suggestion is to use two LM3778s, with a +/- 32Vdc power supply. Rate the power supply for about 100W of DC (or +/- 32V 1.6A). This needs a 42-45Vct transformer rated at 3A or so to be bulletproof. 10,000uF on each DC side should work OK.  Toroidal transformers will be about half the weight. Drive one 12" speaker with each amplifier.

Company named Antek has toroids on ebay - or did at one time.

I can help with suggestions if you get stuck.

[tonyharker]

Do you really mean LM3778 or do you mean LM3886??

[Jaicen_solo]

That was going to be my next question, I can't find any LM3778's in the UK.

[R.G.]

Sorry - LM3886. I was typing while distracted.

[Jaicen_solo]

Thanks R.G, that looks to be exactly what i'm looking for then. I found a project on GGG using this chip, so i'll jump straight in with that one after I've read the datasheets, and take it from there.
Is there anywhere I can find out how much power I can put through a Celestion G12H? They seem to have fairly measly ratings, don't want to blow anything up!

[tonyharker]

30 watts

[R.G.]

Is there anywhere I can find out how much power I can put through a Celestion G12H? They seem to have fairly measly ratings, don't want to blow anything up!

One interesting thing about using +/- 30V to +/- 33V for power supplies is that a solid state amp of typical design only puts out about 30Wrms with this supply if the load is 8 ohms.

Thomas Vox amps used +/-31V supplies on all of their bigger line of amps. The Beatle used +/-31V, and ran four 8 ohm speakers in parallel, which produced 30W from each, for a total of 120Wrms. The Royal Guardsman used the same power supply, and only two 8 ohm speakers in parallel, and got 60W total. The Buckingham and Viscount used the same power supply and two 16 ohm speakers in parallel for 8 ohms, and put out 30Wrms.

So yes, a Celestion 30 tends to die if you drive it with much over 30Wrms. If you have a much higher power amplifier, you have to series/parallel them to use that power without endangering speakers.

I personally like the idea of an easy-to-build 30Wrms power amp driving a 30Wrms speaker, one-to-one, and using enough speaker+amp modules to get as much power as you need. Need 30W? Use one. 100-120W? Use four. 1 KILO-watt? Use 30 of them.  If one dies in a multi-amp/speaker setup, you keep on rolling, just a little quieter.

[Jaicen_solo]

The redundancy you mention is part of the appeal of the design I have in mind.
However, 30w into a 30w speaker seems like its a bit borderline in terms of being bulletproof.
I may try a 2x12 mono setup just to be safe.

[R.G.]

It's only borderline if what you feed it is a 30W sine wave, or pre-compressed music or 30W of white noise. Normal hifi music tends to have a 20db (i.e. 10:1) peak to average power level crest factor. The peaks have to be reproduced cleanly or it will sound ugly, but the average power that heats the speakers is about 1/10 of that. Musical instrument speakers lead a harsher life, but still the ratio of peak (and by that I mean RMS power sustained for more than a few seconds) to average power is between 3:1 and 5:1 unless you're doing something very strange ahead of it.

The bottom line is that a 30Wrms power amp driving a 30Wrms musical instrument speaker is pretty safe. (Note: several of you are twitching to note that distortion can kill tweeters at levels far below max; this is kinda true, but not an issue with a Celestion 30 as in this case.)

I would have no hesitation using a setup like this.

[PRR]

The GUITAR-speaker makers know the users are non-technical. A "30W" will be very hard to kill with a 30 Watt amplifier.

The flip-side of the low ratings on the old-style speakers is that it *may* imply a very low-mass cone/coil assembly and much higher midrange efficiency (more roar for your Watt). OTOH when 100W beasts became common designs had to shift to large heavy coils and efficiency dropped. Small amps often work well with low-rated speakers.

(There are exceptions. You can't in-all-ways beat a JBL 130 or 120 or the EVMs, even though they were designed for LARGE amps. OTOH they are heavy and costly. And perhaps a bit tame when not pushed by a herd of horsepower.)